Ionic Liquid-Induced Local Charge Compensation: Effects on Back

Article pubs.acs.org/JPCC

Ionic Liquid-Induced Local Charge Compensation: Effects on Back Electron-Transfer Rates in Dye-Sensitized TiO2 Thin Films Valeria Saavedra Becerril, Daniele Franchi,† and Maria Abrahamsson* Department of Chemistry and Chemical Engineering, Chalmers University of Technology, Kemivägen 10, 41296 Göteborg, Sweden ABSTRACT: The effect of ionic liquid electrolytes on back electron-transfer rates for dye-sensitized TiO2 thin films was investigated using microsecond−millisecond transient absorption. For D35/TiO2 and [Ru(dcb)3]2+/TiO2 in electrolytes based on 1-alkyl-3-methyl-imidazolium hexafluorophosphate, significantly slower back electron-transfer rates, compared to those observed in neat acetonitrile (CH3CN) and LiClO4/ CH3CN, were found. Surprisingly, no such trends were observed for N3/TiO2 under the same conditions. This inconsistency points to the need for mechanistic understanding of how the structure and properties of dyes affect the electrolyte−dye interface interactions in ionic liquid (IL)based dye-sensitized solar cells (DSSCs). To explain the observed behavior we propose an electrostatic effect at the TiO2−electrolyte interface, where the bulky IL cations rearrange at the TiO2 surface, locally compensating the charge. This would be consistent with N3 behaving differently because of its negatively charged SCN− ligands. This accumulation of cations at the interface affects the interaction between conduction band TiO2 electrons and the oxidized dye. As a result, slower back electron-transfer rates are observed when charge is effectively compensated. Therefore, here, the study of back electron-transfer kinetics was used as an indirect probe of local charge compensation at the dye−semiconductor−electrolyte interfaces. The results show that the mechanism of local charge compensation is dependent on dye structure.

■

INTRODUCTION The overall efficiency of dye-sensitized mesoporous nanocrystalline semiconductor devices for solar energy conversion depends on the individual efficiencies of the various interfacial charge-transfer processes that occur subsequent to light absorption.1−3 The details of photocurrent generation and enhancement are well understood today,4−7 but much insight is still needed before the back electron-transfer reaction, from TiO2 conduction band electrons to the oxidized dye, can be controlled and exploited and before we fully understand the mechanisms behind the electrostatic interactions at the TiO2− dye−electrolyte interface.8 Understanding and controlling back electron transfer is crucial both for open-circuit voltage optimization and solar fuels generation.9,10 Many successful approaches to decrease the rate of back electron transfer are based on dye structure modifications. For example, the use of sensitizers that undergo intramolecular hole-transfer reactions subsequent to electron injection typically results in long-lived charge separation; however, this often occurs at the expense of considerable losses in the free energy stored in the chargeseparated state.11,12 Another strategy that has been pursued is to increase the distance between the TiO2 surface and the chromophore by introduction of spacers; incorporation of phenyleneethynylene spacers can slow back electron transfer without concomitant free energy losses; however, this sometimes occurs at the cost of low injection efficiencies.13,14 In this study we take the simpler approach of changing the © XXXX American Chemical Society

environment at the interface rather than modifying the dye structure, by introducing a pure ionic liquid (IL) electrolyte. This strategy provides the ability to modify the back electrontransfer kinetics and gain insight about the effects that large ionic liquid cations can have at the charged TiO2 interface. Several reports on the effect of ILs on the electron-transfer processes in DSSC have been published;15−18 however, to the best of our knowledge, none of them has solely focused on the back electron-transfer reaction. Characterization of DSSCs with IL electrolytes has shown that electron recombination is accelerated and regeneration is slowed compared to DSSCs with organic solvent-based electrolytes. This is the case for DSSCs using either organic dyes or ruthenium complexes as photosensitizers.16,18,19 The origin of these effects is not yet fully understood, and speculation that diffusion limitations caused by increased solution viscosity have been disproven as the sole reason for the observed behavior.18 There is an apparent need for mechanistic insight, especially because minimizing recombination losses is crucial not only for efficient electricity generation in DSSCs but also for applications where multiple electron transfer is required. Thus, it becomes necessary to expand our understanding of the influence of ILs on the interfacial electron-transfer processes. Received: June 16, 2016 Revised: August 24, 2016

A

DOI: 10.1021/acs.jpcc.6b06088 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C Chart 1. Dyes and IL Cations Used in This Studya

a

From left to right: N3; [Ru(dcb)3]Cl2; D35; and the 1-butyl, 1-hexyl-, and 1-octyl-3-methylimidazolium cations.

Table 1. Maximum Absorption Wavelengths (Nanometers) in Solution and on TiO2 in the Environments Indicated

a

dye

EtOH or MeOHa

TiO2/air

TiO2/CH3CN

TiO2/0.1 M LiClO4

TiO2/BMIMPF6

TiO2/HMIMPF6

TiO2/OMIMPF6

D35 N3 [(Ru(dcb)3)Cl2]

476 533 462

465 524 452

460 534 467

479 527 467

471 524 467

473 525 468

478 525 464

Absorption spectra of D35 and N3 were recorded in EtOH. Absorption spectra of [Ru(dcb)3]2+ were recorded in MeOH.

Figure 1. From left to right, electronic absorption spectra of D35/TiO2, N3/TiO2, and [Ru(dcb)3]2+/TiO2 in air (gray); neat CH3CN (red); 0.1 M LiClO4/CH3CN (yellow); and ILs BMIMPF6 (blue), HMIMPF6 (green), and OMIMPF6 (purple). Spectra have been normalized to a maximum absorption of 1.0 for the dye peak.

mM D35 (Dyenamo), 0.1 mM N3 (Solaronix), or 0.1 mM [(Ru(dcb)3)Cl2] (Solaronix) in ethanol or methanol. Prior to sensitization the films were heated to 80 °C. For spectroscopic measurements, a drop of the electrolytes was added on top of the samples and a microscope glass coverslip was placed on top of them. Surface coverages were determined by measuring the absorption spectrum of the dye bath solution before and after sensitization of the films. Using the molar absorption coefficients of the dyes, the number of molecules on the 1 cm2 film was determined. Steady-State Absorption Spectroscopy. Electronic spectra were recorded using a Varian Cary 50 Bio spectrometer. Transient Absorption Spectroscopy. Single-wavelength transient spectroscopic measurements were performed using a Continuum Surelight II, Nd:YAG laser at 10 Hz repetition rate with a Continuum Surelight optical parameter oscillator, producing pulsed (ca. 8 ns fwhm) 425 nm excitation light. The power of the pulse was adjusted between 1 and 3 mW cm−2 depending on the surface coverage of the samples to obtain a similar density of injected electrons. Probe light was generated with a Quartz Tungsten Halogen lamp (Ushio) powered by a radiometric power supply (Newport). The wavelength of interest was selected by two Cornerstone 130 monochromators (Oriel Instruments). Data was collected using either a five-stage photomultiplier tube, PMT (Applied Photophysics), or a Costronics photodetector amplified by a Costronics Optical Transient Amplifier connected to a

The kinetics of these electron-transfer processes can be measured on dye-sensitized semiconductor thin films rather than full devices to better assess the mechanistics of the back electron-transfer process. In this work we have measured back electron-transfer kinetics using single-wavelength transient absorption spectroscopy in dye−TiO2 samples with ionic liquid electrolytes, where the dyes used in this work were D35, ((E)3-(5-(4-(bis(2′,4′-dibutoxy-[1,1′-biphenyl]-4-yl)amino)phenyl)thiophen-2-yl)-2-cyanoacrylic acid),20 N3 (cis-bis(isothiocyanato)bis(2,2′-bipyridyl-4,4′-dicarboxylato ruthenium(II)),21 and [Ru(dcb)3]Cl2 (dcb is bis(2,2′-bipyridyl-4,4′dicarboxylic acid)22 (Chart 1). The ionic liquids used are three imidazolium-based ILs: 1-butyl, 1-hexyl, and 1-octyl-3methylimidazolium hexafluorophosphate (BMIMPF 6 , HMIMPF6, and OMIMPF6 respectively) (Chart 1). Our results show that IL electrolytes can significantly slow the back electron-transfer rate without considerably affecting electron injection efficiencies; however, this appears to be strongly dependent on the electronic structure of the dye.

■

EXPERIMENTAL SECTION Sample Preparation. Transparent mesoporous nanocrystalline TiO2 films of 1 cm2 area, ca. 6 μm thickness, and average nanoparticle diameter of 20 nm were prepared by doctor-blading technique on FTO (TEC-15) (Solaronix) substrates using commercially available TiO2 paste (18-NR-T, Dyesol). After paste deposition, the samples were progressively heated to 450 °C in air flow. Dye baths contained either 0.3 B


Article

The Journal of Physical Chemistry C Tektronix TDS 2022 oscilloscope. The data was analyzed using the Igor Pro software from WaveMetrics.

■

RESULTS AND DISCUSSION In this work, we chose to work with well-known and wellcharacterized dyes, one organic dye (D35) and two Ru(II)based dyes, N3 and [Ru(dcb)3]2+. D35 was chosen because of its appreciable stability, good device performance, and ultrafast electron injection into TiO2 conduction band within tens of femtoseconds, even in air.23,24 N3 was chosen for similar reasons as D35 while [Ru(dcb)3]2+ was employed as a model dye for all the Ru2+ dyes investigated as sensitizers for DSSCs.25 Neat acetonitrile and 0.1 M LiClO4 in acetonitrile were used as reference electrolytes. Our first transient absorption measurements at the submicrosecond scale, on D35/TiO2 samples, suggested that the back electron-transfer kinetics were significantly affected by the ILs without affecting electron injection. Thereafter, we carefully studied the back electrontransfer kinetics for all the dyes and extracted the corresponding rate constants. Ground-State Absorption Spectroscopy. In Table 1 and Figure 1 we summarize the electronic absorption data of the dye−TiO2 assemblies in contact with the different electrolytes and solvents used. For the Ru-based photosensitizers, very small spectral shifts were observed upon addition of the different solvents, with the exception of N3 in acetonitrile where a more pronounced red-shift is observed. For D35 the absorption in neat acetonitrile is the most blue-shifted with λmax of 460 nm, consistent with previously reported results.24,26 All the other D35/TiO2 samples display bathochromic shifts with respect to acetonitrile, with minor individual differences. This suggests that the ground-state interactions are similar in all electrolytes containing a significant amount of ionic species. We note that the small but significant red-shift of the D35/TiO2 absorption peak upon addition of the ILs increases with the length of the alkyl chain in the alkylimidazolium cations. Nanosecond Transient Absorption Spectroscopy. The interfacial back electron-transfer kinetics were monitored with single-wavelength transient absorption measured on the sensitized thin films. Surface coverages varied between 10−7 and 10−9 mol/cm2 depending on the dye. The laser pulse intensity was adjusted according to the absorbance of the samples at the excitation wavelength. Also, the absorbance of all samples with the same dye was kept as similar as possible. Samples with absorbance variations higher than 10% of the average were not considered for measurement. This compensates for the difference in surface coverages with the purpose of obtaining comparable kinetics for all the three dyes studied. By doing this, we obtain similar density of injected electrons in the conduction band according to eq 1.27 Here, λex is the excitation wavelength, Fex the energy fluence per laser pulse, A the absorbance of the dye at λex, r the average particle radius, d the thickness of the mesoporous film, and φ its porosity. This accounts for the strong dependence of electron dye recombination kinetics upon electron density in the conduction band of the semiconductor.3,28 ⟨x⟩ =

λex ·Fex ·(10−A) ·4π ·r 3 3h·c·d ·φ

Figure 2. Normalized representative individual traces and corresponding fits measured by single-wavelength transient absorption after pulsed 425 nm excitation. The traces were obtained with photodiode detection at 750 nm for D35/TiO2 (top left), 500 nm for [Ru(dcb)3]2+/TiO2 (top right), and 750 nm for N3/TiO2 (bottom) in neat CH3CN (red); 0.1 M LiClO4/CH3CN (yellow); and ILs BMIMPF6 (blue), HMIMPF6 (green), and OMIMPF6 (purple). The inset in the N3/TiO2 graph shows the KWW fit to one of the curves.

strong positive absorption signals. A weak absorption from electrons accumulated in the conduction band could have a very small contribution24 to this signal, but because care was taken to minimize the number of injected electrons this contribution should be negligible. The oxidized form of [Ru(dcb)3]2+ exhibits a very broad and weak absorption band above 600 nm; therefore, the kinetics were probed at 500 nm instead, corresponding to the ground-state bleach. Although electron injection from [Ru(dcb)3]2+ to TiO2 is considerably less efficient than in N3 and D35, we clearly observe the formation of oxidized [Ru(dcb)3]2+ based on the decay time of the ground-state bleach (Figure 2). As can be readily seen in Figure 2, the back electron-transfer kinetics were visibly slower in all the ILs compared to acetonitrile for both D35 and [Ru(dcb)3]2+. In contrast, N3sensitized thin films show no appreciable differences in the kinetics observed no matter the solvent used. We did not observe significant differences in the initial values of ΔA signals, even at the best time resolution used, in any of our samples. Therefore, we do not have reasons to believe that the presence of ionic liquids had a significant effect on the quantum yields of injection because that would typically result in easily detectable differences in ΔA amplitudes.28 The decay traces could not be fitted to a monoexponential decay model; instead, we used the Kohlrausch−Williams− Watts (KWW) kinetic model,25,26 eq 2a,b, or a biexponential decay model, eq 3a,b, for adequate description of the collected recombination data. The rate constants extracted from the various decay models are summarized in Tables 2 and 3. I = (I0 − If )exp[−(kobst )β ] + If

(1)

Figure 2 shows representative back electron-transfer kinetic traces with corresponding fits. The decays of D35 and N3 were probed at 750 nm, where the oxidized forms of both dyes have

⟨τww ⟩ = C

⎛1⎞ Γ⎜ ⎟ kobsβ ⎝ β ⎠

(2a)

1

(2b) DOI: 10.1021/acs.jpcc.6b06088 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C ⎧ −(t − t0) ⎫ ⎧ −(t − t0) ⎫ ⎬ ⎬ + A 2 exp⎨ I = I0 + A1exp⎨ τ1 τ2 ⎭ ⎩ ⎭ ⎩

electrolytes they are more similar, a clear indication that the ILs affect the recombination rates differently depending on dye structure. The small differences between the extracted rate constants for a specific dye in the different ILs provide no evidence for a clear trend with regards to back electron-transfer rate and size/bulkiness of the IL cation. However, we do note that the slowest back electron-transfer for both D35 and [Ru(dcb)3]2+ is observed in the presence of the bulkiest cation, OMIM+. Fitting the [Ru(dcb)3]2+ decay traces with the KWW model resulted in large and nonsystematic variations in β which could not be related to differences between samples; therefore, the model was not considered robust in this case. Instead, the experimental data was well-described by the double-exponential model (eq 3a,b), where τ1 and τ2 are the values of the decay time constants and A1 and A2 their corresponding amplitudes. The extracted decay rate constants, k1 and k2, are calculated as the inverse of the decay time constants obtained from the fits (eq 3b). As mentioned above, the D35 data could be reasonably well reproduced also by a biexponential model, and to enable straightforward comparisons between D35 and [Ru(dcb)3]2+, the D35 data was also fitted to this model. Small differences are found when comparing the values of the rate constants for [Ru(dcb)3]2+/TiO2 in neat acetonitrile and upon addition of the ILs. The differences in the relative amplitudes of k1 ans k2 confirm that back electron transfer is slower in the presence of ILs (Table 3). Summarizing these data, we can conclude that slowest decay is observed with OMIMPF6, for both D35 and [Ru(dcb)3]2+ samples. However, the effect of Li+ is not the same for all the dyes. The data analysis provided above confirms that charge recombination in D35 and [Ru(dcb)3]Cl2 sensitized TiO2 becomes slower in the presence of the ILs studied here. Given the consistency of our data, the 2- to 4-fold change in charge recombination rate constants observed here should be considered a significant change. Because the same was not observed for N3 sensitized films, a detailed discussion about the possible reasons for the observed behavior is warranted. Discussion about the Differences in Recombination Electron-Transfer Rates. The observed decreased interfacial recombination rates in D35/TiO2 and [Ru(dcb)3]2+/TiO2 samples cannot simply be explained by differing concentrations of TiO2 electrons. As outlined above, care was taken to ensure the same concentration of injected electrons in all samples containing the same dye. The concentration of electrons in the conduction band is dependent not only on the irradiation power but also on the electron injection efficiencies for a specific dye. This process is known to be affected by the use of additives in the electrolyte

(3a)

k1 =

1 , τ1

k2 =

1 τ2

(3b)

Table 2. Representative Back Electron-Transfer Rate Constants, kcr, and β Parameters Obtained from KWW Model Fitting of Transient Absorption Data of D35/TiO2 and N3/TiO2 Samples in the Electrolyte Indicated D35 electrolyte CH3CN LiClO4 BMIMPF6 HMIMPF6 OMIMPF6

3

−1 a

± ± ± ± ±

1.3 8.8 2.7 2.2 0.7

N3 β

kcr (10 s ) 11.2 26.1 4.3 7.5 2.8

0.42 0.42 0.41 0.41 0.48

a

± ± ± ± ±

3

−1 a

± ± ± ± ±

2.4 3.2 5.4 3.1 5.3

βa

kcr (10 s ) 0.01 0.08 0.07 0.06 0.09

12.6 20.2 15.3 16.8 18.3

0.24 0.24 0.29 0.24 0.30

± ± ± ± ±

0.02 0.02 0.06 0.05 0.02

a

The indicated uncertainties represent the standard deviation obtained from averaging individual fits of single-wavelength transient absorption decays from triplicate samples.

In eq 2a, I0 and If are the initial and final amplitudes, respectively, and β is related to the width of a stretched Levy distribution of rate constants, taking values between zero and unity.13,29 The first moment of the KWW function is given by eq 2b, where Γ is the gamma function.30 Representative back electron-transfer rate constants, kcr, were taken as the reciprocal of ⟨τww⟩. The N3/TiO2 samples were all better described by the KWW model, and attempts to use the biexponential model yielded nonsensical output. In contrast, D35/TiO2 decay traces could be satisfactorily fitted to both models but were better described by the KWW model. A detailed analysis of the obtained β values suggests no significant difference in the distribution of rate constants for a given dye, which makes comparison of the obtained rate constants straightforward and provides evidence that the kinetic model is robust. However, the β-values for N3, around 0.25, are lower than the β-values for D35, indicating a slightly different distribution of rate constants and consequently different interfacial properties in the two cases. The extracted rate constants also confirm the visual inspection, that charge recombination in IL electrolytes is slowed by at least a factor of 4 and 2 compared to LiClO4 and neat acetonitrile, respectively. Furthermore, the extracted rate constants confirm that the recombination is faster in N3 compared to D35, in all ILs, while in acetonitrile-based

Table 3. Representative Back Electron-Transfer Rate Constants Obtained from Double Exponential Fitting of Transient Absorption Data of D35/TiO2 and [Ru(dcb)3]2+/TiO2 in the Electrolytes Indicated [Ru(dcb)3]2+

D35 electrolyte

A1 %

A2 %

CH3CN LiClO4 BMIMPF6 HMIMPF6 OMIMPF6

43 50 38 44 37

57 50 62 56 63

k1 (104 s−1)a 1.4 2.8 1.3 1.7 0.99

± ± ± ± ±

0.43 1.0 0.5 0.51 0.74

k2 (103 s−1)a

A1 (%)

A2 (%)

± ± ± ± ±

68 46 58 55 52

32 54 42 45 48

1.9 2.8 1.3 1.4 0.92

0.28 0.53 0.28 0.16 0.44

k1 (104 s−1) 4.1 1.2 3.3 2.1 1.6

± ± ± ± ±

0.51 0.09 0.29 0.2 0.33

k2 (103 s−1) 1.5 0.98 2.2 1.3 1.1

± ± ± ± ±

0.18 0.13 0.73 0.15 0.05

a The indicated uncertainties in k1 and k2 represent the standard deviation obtained from averaging individual fits of single-wavelength transient absorption decays from triplicate samples.

D


Article

The Journal of Physical Chemistry C such as Li+ and tert-butylpyridine (tBP), by either shifting the TiO2 conduction band energetics28,31−34 or by surface interactions caused by adsorption of molecules at the TiO2 surface.32,35,36 Given the experimental conditions of our study, the differences in back electron-transfer rates observed with different IL cations cannot be attributed to a conduction band shift. To the best of our knowledge, the few reports regarding conduction band energy shifts in DSSCs with IL electrolytes either involve Li+37 or are anion-correlated.38 Moreover, a general conduction band edge shift caused by IL species should similarly affect the interfacial electron dynamics for all the dyes because the generality of this effect has been confirmed through years of observations.28,31−36 Addition of Li+ had a similar effect on both D35 and N3, where approximately a 2-fold increase in kcr is observed in comparison to neat acetonitrile. Thus, a Li+ induced conduction band edge shift could at least explain the faster kinetics observed for D35/TiO2 and N3/TiO2 in LiClO4/CH3CN, assuming that the charge recombination occurs in the Marcus inverted region.39,40 From this, it follows that a conduction band edge shift would produce an effect of similar magnitude and direction in both D35/TiO2 and N3/ TiO2 samples. Thus, a conduction band edge shift effect cannot justify the slower recombination observed in D35/TiO2 with ILs, given that the effect of ILs in the kcr of D35/TiO2 and N3/ TiO2 is opposite. The slower recombination kinetics in [Ru(dcb)3]2+/TiO2 with ILs compared to that in neat acetonitrile suggests that the electronic structure of the dye plays an important role. However, the effect of Li+ observed for [Ru(dcb)3]2+/TiO2 is somewhat surprising, and speculation about the reasons for this is beyond the scope of this paper. The general high viscosity of IL electrolytes is directly related to changes in the electron dynamics in IL-DSSCs.18,41 Although ionic diffusion is not a limiting process for interfacial charge recombination, we still wanted to make sure that our observations would not have any direct relationship with the general physical properties of the electrolyte. For this purpose, we also measured the back electron-transfer kinetics of D35/ TiO2 thin films in a polymer-based gel with molar composition of 0.1 polyacrylonitrile (PAN), 0.41 ethyl carbonate, and 0.38 propyl carbonate. The viscosity of this gel is at least 3 orders of magnitude higher than that of the ILs.42−44 The recombination kinetics in the polymer are very similar to those in neat acetonitrile (Figure 3). Therefore, we can confirm that the rheological properties of the electrolytes do not affect charge recombination dynamics.

Following that result, we wanted to ensure that pure IL electrolytes do not generally prevent reactivity; therefore, the decay of D35/TiO2 samples was measured using BMIM+ with the noninnocent anions I− and SCN− (Figure 3). Because both I− and SCN− are known to regenerate the oxidized dye, and I− more efficiently than SCN−, faster decay of the oxidized dye signal is expected.45−47 This was indeed observed; in fact, transient absorption kinetics in BMIMI suggested that regeneration of D35 was complete in 2 μs (Figure 3). A bleaching of the ground-state absorption of the dye is observed after the regeneration reaction is finished, which Wang et al. have previously attributed to formation of the reduce dye formed by the reaction of the excited dye with the high concentration of iodide anions.47 Thus, it can be safely concluded that charge recombination becomes slower in the presence of the ILs studied here, when the anions are innocent such as PF6−. This would point to an interfacial effect where the bulky, asymmetric, and highly available cations are rearranged and electroadsorbed at the charged TiO2 surface modifying the interaction between CB electrons and oxidized dye molecules. Considering the size of the cations and the low surface coverage of our sample, it is likely that the cations penetrate the dye layer and adsorb at free TiO2 sites. Compared to the small, solvated lithium ions in organic solvent electrolytes, bulky cations and molecules should be more prone to form layered, compact structures at the charged interface.48 Charge screening due to the creation of a bulky insulating network on the TiO2 surface has been suggested before, with the use of chenodeoxycholic acid as a surface coadsorbent.49 As a result, a change in electronic coupling between the electrons in the conduction band and the oxidized dye molecules is expected.32 This affects the dynamics of back electron transfer. When Meyer and collaborators first demonstrated Stark effects at dye-sensitized TiO2, they predicted that screening of electric fields by chemical species should have an influence on the lifetime of the charge-separated state.50 Our results are in agreement with this. Ionic rearrangement in DSSCs has been suggested to take place on time scales ranging from 10−6 to 10−5 seconds51 and has been experimentally confirmed by Stark effect dynamic studies.52 Thus, a correlation between back electron-transfer dynamics and local charge compensation effect is a reasonable way to explain the observed results. Electrosorption of cations at the TiO2 surface to compensate the charge at the TiO2 surface has been reported before, and although the exact time that cations take to locally reorganize at the charged TiO2 surface is unknown, it is reported that they can remain for a time longer than typical interfacial recombination times before the electrical field is reversed.8 Possible steric and/or electrostatic effects of the IL cations at the TiO2 surface might be very similar for the D35 and [Ru(dcb)3]2+ compounds, neither of which contain charged ligands; therefore, smaller interactions with the IL could be expected. However, in N3, it is possible that the IL cations will have stronger interactions around the SCN− ligands rather than at the charged TiO2 interface, making the accumulation of cations at the TiO2 surface less efficient. To explain the behavior for N3/TiO2 samples, a deeper analysis is needed. The structure of the dye as well as how it arranges at the TiO2 surface can possibly prevent the IL cations from penetrating the surface. Therefore, aggregation and πstacking effects should be taken into consideration. It is known

Figure 3. Left: Individual traces measured by single-wavelength transient absorption at 750 nm (photodiode detection) after pulsed 425 nm excitation of D35/TiO2 thin films with the various solvents. Right: Individual traces measured by single-wavelength transient absorption after pulsed 425 nm excitation, measured at 600 nm (PMT detection) for D35/TiO2 thin films with the various electrolytes. E


The Journal of Physical Chemistry C

■

ACKNOWLEDGMENTS The Swedish Research Council (Vetenskapsrådet) is acknowledged for financial support. D.F. thanks the Foundation BLANCEFLOR Boncompagni Ludovisi, née Bildt for a personal scholarship. We also thank Dr. Anna Martinelli for providing the HMIMPF6, Erik Anvell for preparing the polymer electrolytes, and Dr. Maurizio Furlani for valuable discussions.

that planar dyes, and specifically D−π−A systems such as the organic D35, are much more susceptible to aggregate and form π-stacks than pseudo-octahedral dyes such as N3.53−55 Therefore, aggregation effects are more likely to take place in D35 rather than N3 samples, considering also the higher surface coverage obtained in D35/TiO2 samples. Thus, aggregation is not likely to be the reason why we did not observe the same effect in D35 and N3 samples with ILs. Both N3 and [Ru(dcb)3]2+ are pseudo-octahedral complexes, but their ligands have very different electronic properties, and a detailed comparison might provide useful insight. The observed behavior could be due to the presence of negatively charged SCN− ligands in N3, which would be in agreement with the hypothesis of an electrostatic effect at the dye−TiO2− electrolyte interface being the main reason for the observed behavior. In the above analysis, we have provided arguments to explain the unexpected dye-structure-dependent dynamics of charge recombination observed in this study. Effects from differences in injection efficiency, aggregation, viscosity, dye geometry, and secondary chemical reactions are dismissed both from our experimental observations and literature findings. Thus, an electrostatic effect at the interface is considered to explain our observations.

■

ABBREVIATIONS IL, ionic liquid; BMIMPF6, 1-butyl-3-methylimidazolium hexafluorophosphate; HMIMPF6, 1-hexyl-3-methylimidazolium hexafluorophosphate; OMIMPF6, 1-butyl-3-methylimidazolium hexafluorophosphate; BMIMSCN, 1-butyl-3-methylimidazolium thiocyanate; BMIMI, 1-butyl-3-methylimidazolium iodide; CH3CN, acetonitrile; LiClO4, lithium perchlorate

■

REFERENCES

(1) Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. DyeSensitized Solar Cells. Chem. Rev. 2010, 110, 6595−6663. (2) O’Regan, B. C.; Durrant, J. R. Kinetic and Energetic Paradigms for Dye-Sensitized Solar Cells: Moving from the Ideal to the Real. Acc. Chem. Res. 2009, 42 (11), 1799−1808. (3) Listorti, A.; O’Regan, B.; Durrant, J. R. Electron Transfer Dynamics in Dye-Sensitized Solar Cells. Chem. Mater. 2011, 23 (15), 3381−3399. (4) Feldt, S. M.; Gibson, E. A.; Gabrielsson, E.; Sun, L.; Boschloo, G.; Hagfeldt, A. Design of Organic Dyes and Cobalt Polypyridine Redox Mediators for High-Efficiency Dye-Sensitized Solar Cells. J. Am. Chem. Soc. 2010, 132 (46), 16714−16724. (5) Albero, J.; Atienzar, P.; Corma, A.; Garcia, H. Efficiency Records in Mesoscopic Dye-Sensitized Solar Cells. Chem. Rec. 2015, 15 (4), 803−828. (6) Qi, J. F.; Dang, X. N.; Hammond, P. T.; Belcher, A. M. Highly Efficient Plasmon-Enhanced Dye-Sensitized Solar Cells through Metal@Oxide Core-Shell Nanostructure. ACS Nano 2011, 5 (9), 7108−7116. (7) Wang, H.; Wang, B. Y.; Yu, J. C.; Hu, Y. X.; Xia, C.; Zhang, J.; Liu, R. Significant Enhancement of Power Conversion Efficiency for Dye Sensitized Solar Cell Using 1D/3D Network Nanostructures as Photoanodes. Sci. Rep. 2015, 5, 9305. (8) Yang, W.; Pazoki, M.; Eriksson, A. I. K.; Hao, Y.; Boschloo, G. A Key Discovery at the TiO2/Dye/Electrolyte Interface: Slow Local Charge Compensation and a Reversible Electric Field. Phys. Chem. Chem. Phys. 2015, 17 (26), 16744−16751. (9) Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. DyeSensitized Solar Cells. Chem. Rev. 2010, 110, 6595−6663. (10) Hammarström, L. Accumulative Charge Separation for Solar Fuels Production: Coupling Light-Induced Single Electron Transfer to Multielectron Catalysis. Acc. Chem. Res. 2015, 48, 840−85. (11) Abrahamsson, M.; Hedberg, J. H. J.; Becker, H. C.; Staniszewski, A.; Pearson, W. H.; Heuer, W. B.; Meyer, G. J. High Extinction Coefficient Ru-Sensitizers that Promote Hole Transfer on Nanocrystalline TiO2. ChemPhysChem 2014, 15 (6), 1154−1163. (12) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.; Meyer, G. J. Long-Lived Photoinduced Charge Separation across Nanocrystalline TiO2 Interfaces. J. Am. Chem. Soc. 1995, 117, 11815− 11816. (13) Abrahamsson, M.; Johansson, P. G.; Ardo, S.; Kopecky, A.; Galoppini, E.; Meyer, G. J. Decreased Interfacial Charge Recombination Rate Constants with N3-type Sensitizers. J. Phys. Chem. Lett. 2010, 1, 1725−1728. (14) Johansson, P. G.; Zhang, Y.; Meyer, G. J.; Galoppini, E. Homoleptic ″Star″ Ru(II) Polypyridil Complexes: Shielded Chromophores to Study Charge-Transfer at the Sensitizer-TiO2 Interface. Inorg. Chem. 2013, 52 (14), 7947−7957.

■

CONCLUSIONS We have shown that ionic liquids can be used to decrease the interfacial charge recombination between TiO2 electrons and oxidized D35, as well as between TiO2 electrons and [Ru(dcb)3]2+ while the recombination rates in N3/TiO2 assemblies remained unaffected. By careful examination of the obtained kinetic data and literature findings we attribute this behavior to an electrostatic effect at the TiO2−dye−electrolyte interface. A mechanism where cations in the ILs are accumulated and/or adsorbed at the charged TiO2 surface to locally compensate the charge is proposed. Electrostatic hindrance results in changes in the recombination rates. We note that the extent of this effect might depend on the size of the ionic liquid cations and the structure of the dye, being the most important aspect the charge on the ligands. A consequence of the findings presented here is that the dye structure should be considered not only for its electron injection properties but also for how it may interact with the surrounding electrolyte. The study of electron interfacial dynamics can be used for probing local charge compensation and similar phenomena. In addition, control of interfacial charge recombination by modification of the surrounding electrolyte seems to be a practical and attractive strategy that could be useful in many solar energy conversion applications.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +46317723069. Present Address †

D.F.: Departimento di Chimica “Ugo Schiff”, Universita degli Studi di Firenze, Via della Lastruccia 13, 50019 Sesto Fiorentino (FI), Italy.

Notes

The authors declare no competing financial interest. F


Article

The Journal of Physical Chemistry C

(33) Yu, Q. J.; Wang, Y. H.; Yi, Z. H.; Zu, N. N.; Zhang, J.; Zhang, M.; Wang, P. High-Efficiency Dye-Sensitized Solar Cells: The Influence of Lithium Ions on Exciton Dissociation, Charge Recombination, and Surface States. ACS Nano 2010, 4 (10), 6032− 6038. (34) Koops, S. E.; O’Regan, B. C.; Barnes, P. R.; Durrant, J. R. Parameters Influencing the Efficiency of Electron Injection in DyeSensitized Solar Cells. J. Am. Chem. Soc. 2009, 131 (13), 4808−18. (35) Boschloo, G.; Häggman, L.; Hagfeldt, A. Quantification of the Effect of 4-tert-Butylpyridine Addition to I-/I3- Redox Electrolytes in Dye-Sensitized Nanostructured TiO2 Solar Cells. J. Phys. Chem. B 2006, 110 (26), 13144−131150. (36) Gothelid, M.; Yu, S.; Ahmadi, S.; Sun, C. H.; Zuleta, M. Structure-Dependent 4-Tert-Butyl Pyridine-Induced Band Bending at TiO2 Surfaces. Int. J. Photoenergy 2011, 2011, 1. (37) Bai, Y.; Zhang, J.; Wang, Y. H.; Zhang, M.; Wang, P. LithiumModulated Conduction Band Edge Shifts and Charge-Transfer Dynamics in Dye-Sensitized Solar Cells Based on a Dicyanamide Ionic Liquid. Langmuir 2011, 27 (8), 4749−4755. (38) Zhang, M.; Zhang, J.; Bai, Y.; Wang, Y. H.; Su, M.; Wang, P. Anion-Correlated Conduction Band Edge Shifts and Charge Transfer Kinetics in Dye-Sensitized Solar Cells with Ionic Liquid Electrolytes. Phys. Chem. Chem. Phys. 2011, 13 (9), 3788−3794. (39) Moser, J. E.; Gratzel, M. Observation of Temperature Independent Heterogeneous Electron-Transfer Reactions in the Inverted Marcus Region. Chem. Phys. 1993, 176 (2−3), 493−500. (40) Kuciauskas, D.; Freund, M. S.; Gray, H. B.; Winkler, J. R.; Lewis, N. S. Electron Transfer Dynamics in Nanocrystalline Titanium Dioxide Solar Cells Sensitized with Ruthenium or Osmium Polypyridyl Complexes. J. Phys. Chem. B 2001, 105 (2), 392−403. (41) Cao, Y. M.; Zhang, J.; Bai, Y.; Li, R. Z.; Zakeeruddin, S. M.; Gratzel, M.; Wang, P. Dye-Sensitized Solar Cells with Solvent-Free Ionic Liquid Electrolytes. J. Phys. Chem. C 2008, 112 (35), 13775− 13781. (42) Geng, Y.; Chen, S.; Wang, T.; Yu, D.; Peng, C.; Liu, H.; Hu, Y. Density, Viscosity and Electrical Conductivity of 1-butyl-3-methylimidazolium hexafluorophosphate + monoethanolamine and + N, Ndimethylethanolamine. J. Mol. Liq. 2008, 143 (2−3), 100−108. (43) Harris, K. R.; Kanakubo, M.; Woolf, L. Temperature and Pressure Dependence of the Viscosity of the Ionic Liquids 1-Hexyl-3methylimidazolium Hexafluorophosphate and 1-Butyl-3-methylimidazolium Bis(trifluoromethylsulfonyl)imide. J. Chem. Eng. Data 2007, 52 (3), 1080−1085. (44) Dintcheva, N.; Furlani, M.; Jayasundara, W.; Bandara, L.; Mellander, B. E.; La Mantia, F. Rheological Behavior of PAN-based Electrolytic Gel Containing Tetrahexylammonium and Magnesium Iodide for Photoelectrochemical Applications. Rheol. Acta 2013, 52, 881−889. (45) Oskam, G.; Bergeron, B. V.; Meyer, G. J.; Searson, P. C. Pseudohalogens for Dye-Sensitized TiO2 Photoelectrochemical Cells. J. Phys. Chem. B 2001, 105, 6867−6873. (46) Wang, P.; Zakeeruddin, S. M.; Moser, J.-E.; Humphry-Baker, R.; Grätzel, M. A Solvent-Free, SeCN−/(SeCN)3− Based Ionic Liquid Electrolyte for High-Efficiency Dye-Sensitized Nanocrystalline Solar Cells. J. Am. Chem. Soc. 2004, 126, 7164−7165. (47) Wang, P.; Wenger, B.; Humphry-Baker, R.; Moser, J. E.; Teuscher, J.; Kantlehner, W.; Mezger, J.; Stoyanov, E. V.; Zakeeruddin, S. M.; Gratzel, M. Charge Separation and Efficient Light Energy Conversion in Sensitized Mesoscopic Solar Cells Based on Binary Ionic Liquids. J. Am. Chem. Soc. 2005, 127 (18), 6850−6856. (48) Ivanistsev, V.; Federov, M. V. Interfaces Between Charged Surfaces and Ionic Liquids: Insights from Molecular Simulations. The Electrochemical Society Interface. Spring 2014, 65−69. (49) Salvatori, P.; Marotta, G.; Cinti, A.; Anselmi, C.; Mosconi, E.; De Angelis, F. Supramolecular Interactions of Chenodeoxycholic Acid Increase the Efficiency of Dye-Sensitized Solar Cells Based on a Cobalt Electrolyte. J. Phys. Chem. C 2013, 117, 3874−3887. (50) Ardo, S.; Sun, Y.; Staniszewski, A.; Castellano, F. N.; Meyer, G. J. Stark Effects after Excited-State Interfacial Electron Transfer at

(15) Paulsson, H.; Kloo, L.; Hagfeldt, A.; Boschloo, G. Electron Transport and Recombination in Dye-Sensitized Solar Cells With Ionic Liquid Electrolytes. J. Electroanal. Chem. 2006, 586, 56−61. (16) Gorlov, M.; Kloo, L. Ionic liquid Electrolytes for Dye-Sensitized Solar Cells. Dalton Trans. 2008, 2655−2666. (17) Mahanta, S.; Furube, A.; Matsuzaki, H.; Murakami, T.; Matsumoto, H. Electron Injection Efficiency in Ru-Dye Sensitized TiO2 in the Presence of Room Temperature Ionic Liquid Solvents Probed by Femtosecond Transient Absorption Spectroscopy: Effect of Varying Anions. J. Phys. Chem. C 2012, 116, 20213−20219. (18) Li, F.; Jennings, J. R.; Wang, X.; Fan, L.; Koh, Z. Y.; Yu, H.; Yan, L.; Wang, Q. Influence of Ionic Liquid on Recombination and Regeneration Kinetics in Dye-Sensitized Solar Cells. J. Phys. Chem. C 2014, 118, 17153−17159. (19) Fabregat-Santiago, F.; Bisquert, J.; Palomares, E.; Otero, L.; Kuang, D.; Zakeeruddin, S. M.; Grätzel, M. Correlation Between Photovoltaic Performance and Impedance Spectroscopy of DyeSensitized Solar Cells Based on Ionic Liquids. J. Phys. Chem. C 2007, 111, 6550−6560. (20) Jiang, X.; Karlsson, K. M.; Gabrielsson, E.; Johansson, E. M. J.; Quintana, M.; Karlsson, M.; Sun, L.; Boschloo, G.; Hagfeldt, A. Highly Efficient Solid-State Dye-Sensitized Solar Cells Based on Triphenylamine Dyes. Adv. Funct. Mater. 2011, 21 (15), 2944−2952. (21) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphrybaker, R.; Muller, E.; Liska, P.; Vlachopoulos, N.; Gratzel, M. Conversion of Light to Electricity by cis-X2bis(2,2′-bipyridyl-4,4′-dicarboxylate)ruthenium(II) Charge-Transfer Sensitizers (X = Cl−, Br−, I−, CN−, and SCN−) on Nanocrystalline Titanium Dioxide Electrodes. J. Am. Chem. Soc. 1993, 115 (14), 6382−6390. (22) Schwarz, O.; van Loyen, D.; Jockusch, S.; Turro, N. J.; Durr, H. Preparation and Application of New Ruthenium(II) Polypyridyl Complexes as Sensitizers for Nanocrystalline TiO2. J. Photochem. Photobiol., A 2000, 132 (1−2), 91−98. (23) Feldt, S. M.; Lohse, P. W.; Kessler, F.; Nazeeruddin, M. K.; Gratzel, M.; Boschloo, G.; Hagfeldt, A. Regeneration and Recombination Kinetics in Cobalt Polypyridine Based Dye-Sensitized Solar Cells, Explained Using Marcus Theory. Phys. Chem. Chem. Phys. 2013, 15 (19), 7087−97. (24) Oum, K.; Lohse, P. W.; Klein, J. R.; Flender, O.; Scholz, M.; Hagfeldt, A.; Boschloo, G.; Lenzer, T. Photoinduced Ultrafast Dynamics of the Triphenylamine-Based Organic Sensitizer D35 on TiO2, ZrO2 and in Acetonitrile. Phys. Chem. Chem. Phys. 2013, 15, 3906−3916. (25) Ardo, S.; Meyer, G. J. Photodriven Heterogeneous Charge Transfer with Transition-Metal Compounds Anchored to TiO2 Semiconductor Surfaces. Chem. Soc. Rev. 2009, 38 (1), 115−164. (26) Pazoki, M.; Lohse, P. W.; Taghavinia, N.; Hagfeldt, A.; Boschloo, G. The Effect of Dye Coverage on the Performance of DyeSensitized Solar Cells with a Cobalt-Based Electrolyte. Phys. Chem. Chem. Phys. 2014, 16, 8503−8508. (27) Kalyanasundaram, K. Dye-sensitized Solar Cells. EPFL Press: Lausanne, Switzerland, 2010. (28) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Stipkala, J. M.; Meyer, G. J. Cation-Controlled Interfacial Charge Injection in Sensitized Nanocrystalline TiO2. Langmuir 1999, 15 (20), 7047−7054. (29) Williams, G.; Watts, D. C. Non-symmetrical Dielectric Relaxation Behaviour Arising from a Simple Empirical Decay Function. Trans. Faraday Soc. 1970, 66, 80−85. (30) Lindsey, C. P.; Patterson, G. D. Detailed Comparison of the Williams-Watts and Cole-Davidson Functions. J. Chem. Phys. 1980, 73, 3348−3357. (31) Nakade, S.; Kanzaki, T.; Kubo, W.; Kitamura, T.; Wada, Y.; Yanagida, S. Role of Electrolytes on Charge Recombination in Dyesensitized TiO2 solar cell (1): The case of solar cells using the I-/I3(−) redox couple. J. Phys. Chem. B 2005, 109 (8), 3480−3487. (32) Katoh, R.; Kasuya, M.; Kodate, S.; Furube, A.; Fuke, N.; Koide, N. Effects of 4-tert-Butylpyridine and Li Ions on Photoinduced Electron Injection Efficiency in Black-Dye-Sensitized Nanocrystalline TiO2 Films. J. Phys. Chem. C 2009, 113 (48), 20738−20744. G


Article

The Journal of Physical Chemistry C Sensitized TiO2 Nanocrystallites. J. Am. Chem. Soc. 2010, 132 (19), 6696−6709. (51) Kuwahara, S.; Taya, S.; Osada, N.; Shen, Q.; Toyoda, T.; Katayama, K. Effect of Electrolyte Constituents on the Motion of Ionic Species and Recombination Kinetics in Dye-Sensitized Solar Cells. Phys. Chem. Chem. Phys. 2014, 16 (11), 5242−5249. (52) Pazoki, M.; Boschloo, G.; Hagfeldt, A. Stark effects in D35Sensitized Mesoporous TiO2: Influence of Dye Coverage and Electrolyte Composition. Electrochim. Acta 2015, 179, 174. (53) Ooyama, Y.; Harima, Y. Molecular Designs and Syntheses of Organic Dyes for Dye-Sensitized Solar Cells. Eur. J. Org. Chem. 2009, 2009 (18), 2903−2934. (54) Ning, Z. J.; Zhang, Q.; Wu, W. J.; Pei, H. C.; Liu, B.; Tian, H. Starburst Triarylamine Based Dyes for Efficient Dye-Sensitized Solar Cells. J. Org. Chem. 2008, 73 (10), 3791−3797. (55) Wang, Z. S.; Koumura, N.; Cui, Y.; Takahashi, M.; Sekiguchi, H.; Mori, A.; Kubo, T.; Furube, A.; Hara, K. HexylthiopheneFunctionalized Carbazole Dyes for Efficient Molecular Photovoltaics: Tuning of Solar-Cell Performance by Structural Modification. Chem. Mater. 2008, 20 (12), 3993−4003.

H


Ionic Liquid-Induced Local Charge Compensation: Effects on Back

Recommend Documents